minifs-v2

#include <linux/module.h>
#include <linux/fs.h>
#include <linux/init.h>
#include <linux/uaccess.h>
#include <linux/slab.h>
#include <linux/statfs.h>
#include <linux/string.h>
#include <linux/list.h>
#include <linux/spinlock.h>

#define MINIFS_NAME "minifs"
#define MINIFS_MAGIC 0x4D494E49 /* "MINI" in ASCII */

/* 文件节点结构 */
struct minifs_inode {
    struct inode vfs_inode;
    char *data;      /* 文件内容 */
    size_t size;     /* 文件大小 */
};

/* 文件元数据（用于动态文件管理） */
struct minifs_file {
    char *name;      /* 文件名 */
    struct inode *inode; /* 关联的 inode */
    struct list_head list; /* 链表节点 */
};

/* 超级块结构 */
struct minifs_sb_info {
    struct super_block *sb;
    struct list_head files; /* 动态文件列表 */
    spinlock_t lock;        /* 保护文件列表 */
    unsigned long next_ino; /* 下一个 inode 编号 */
};

/* 转换为 minifs_inode */
static inline struct minifs_inode *MINIFS_INODE(struct inode *inode)
{
    return container_of(inode, struct minifs_inode, vfs_inode);
}

/* 文件操作 */
static ssize_t minifs_file_read(struct file *file, char __user *buf, size_t count, loff_t *ppos)
{
    struct inode *inode = file_inode(file);
    struct minifs_inode *minifs_inode = MINIFS_INODE(inode);
    size_t len = minifs_inode->size;
    loff_t pos = *ppos;

    if (pos >= len)
        return 0; /* EOF */
    if (pos + count > len)
        count = len - pos;

    if (copy_to_user(buf, minifs_inode->data + pos, count))
        return -EFAULT;

    *ppos = pos + count;
    return count;
}

static ssize_t minifs_file_write(struct file *file, const char __user *buf, size_t count, loff_t *ppos)
{
    struct inode *inode = file_inode(file);
    struct minifs_inode *minifs_inode = MINIFS_INODE(inode);
    loff_t pos = *ppos;
    char *new_data;
    size_t new_size;

    new_size = pos + count > minifs_inode->size ? pos + count : minifs_inode->size;
    new_data = kmalloc(new_size, GFP_KERNEL);
    if (!new_data)
        return -ENOMEM;

    if (minifs_inode->data)
        memcpy(new_data, minifs_inode->data, minifs_inode->size);

    if (copy_from_user(new_data + pos, buf, count)) {
        kfree(new_data);
        return -EFAULT;
    }

    kfree(minifs_inode->data);
    minifs_inode->data = new_data;
    minifs_inode->size = new_size;
    inode->i_size = new_size;
    inode->i_mtime = inode->i_ctime = current_time(inode);

    *ppos = pos + count;
    return count;
}

static const struct file_operations minifs_file_ops = {
    .read = minifs_file_read,
    .write = minifs_file_write,
    .llseek = generic_file_llseek,
};

/* 目录操作 */
static int minifs_iterate(struct file *file, struct dir_context *ctx)
{
    struct inode *inode = file_inode(file);
    struct super_block *sb = inode->i_sb;
    struct minifs_sb_info *sbi = sb->s_fs_info;
    struct minifs_file *file_entry;
    int pos = ctx->pos;

    if (pos == 0) {
        dir_emit(ctx, ".", 1, inode->i_ino, DT_DIR);
        ctx->pos = 1;
        return 0;
    }
    if (pos == 1) {
        dir_emit(ctx, "..", 2, inode->i_ino, DT_DIR);
        ctx->pos = 2;
        return 0;
    }

    spin_lock(&sbi->lock);
    list_for_each_entry(file_entry, &sbi->files, list) {
        if (pos - 2 == 0) {
            dir_emit(ctx, file_entry->name, strlen(file_entry->name),
                     file_entry->inode->i_ino, DT_REG);
            ctx->pos++;
            spin_unlock(&sbi->lock);
            return 0;
        }
        pos--;
    }
    spin_unlock(&sbi->lock);

    return 0;
}

static const struct file_operations minifs_dir_ops = {
    .iterate = minifs_iterate,
};

/* Inode 操作：前置声明 */
static struct dentry *minifs_lookup(struct inode *dir, struct dentry *dentry, unsigned int flags);
static int minifs_create(struct inode *dir, struct dentry *dentry, umode_t mode, bool excl);
static int minifs_unlink(struct inode *dir, struct dentry *dentry);

static const struct inode_operations minifs_inode_ops = {
    .lookup = minifs_lookup,
    .create = minifs_create,
    .unlink = minifs_unlink,
};

/* 超级块操作 */
static const struct super_operations minifs_sb_ops = {
    .statfs = simple_statfs,
    .drop_inode = generic_delete_inode,
};

/* Inode 操作 */
static struct inode *minifs_get_inode(struct super_block *sb, int ino, umode_t mode)
{
    struct inode *inode = new_inode(sb);
    struct minifs_inode *minifs_inode;

    if (!inode)
        return ERR_PTR(-ENOMEM);

    minifs_inode = MINIFS_INODE(inode);
    inode->i_ino = ino;
    inode->i_mode = mode;
    inode->i_atime = inode->i_mtime = inode->i_ctime = current_time(inode);

    if (S_ISREG(mode)) {
        inode->i_fop = &minifs_file_ops;
        minifs_inode->data = NULL;
        minifs_inode->size = 0;
        inode->i_size = 0;
    } else if (S_ISDIR(mode)) {
        inode->i_fop = &minifs_dir_ops;
        inode->i_op = &minifs_inode_ops;
        inode->i_size = 4096;
    }

    return inode;
}

/* 超级块操作 */
static int minifs_fill_super(struct super_block *sb, void *data, int silent)
{
    struct minifs_sb_info *sbi;
    struct inode *root_inode;
    struct minifs_file *file_entry;

    sbi = kzalloc(sizeof(struct minifs_sb_info), GFP_KERNEL);
    if (!sbi)
        return -ENOMEM;

    sb->s_fs_info = sbi;
    sbi->sb = sb;
    sbi->next_ino = 2; /* 1 留给根目录 */
    INIT_LIST_HEAD(&sbi->files);
    spin_lock_init(&sbi->lock);

    sb->s_magic = MINIFS_MAGIC;
    sb->s_blocksize = 1024;
    sb->s_blocksize_bits = 10;
    sb->s_op = &minifs_sb_ops;

    root_inode = minifs_get_inode(sb, 1, S_IFDIR | 0755);
    if (IS_ERR(root_inode)) {
        kfree(sbi);
        return PTR_ERR(root_inode);
    }

    sb->s_root = d_make_root(root_inode);
    if (!sb->s_root) {
        kfree(sbi);
        return -ENOMEM;
    }

    /* 初始化两个默认文件 */
    file_entry = kzalloc(sizeof(struct minifs_file), GFP_KERNEL);
    if (!file_entry)
        return -ENOMEM;
    file_entry->name = kstrdup("hello.txt", GFP_KERNEL);
    file_entry->inode = minifs_get_inode(sb, sbi->next_ino++, S_IFREG | 0644);
    if (IS_ERR(file_entry->inode)) {
        kfree(file_entry->name);
        kfree(file_entry);
        return PTR_ERR(file_entry->inode);
    }
    MINIFS_INODE(file_entry->inode)->data = kstrdup("Hello, World!\n", GFP_KERNEL);
    MINIFS_INODE(file_entry->inode)->size = strlen("Hello, World!\n");
    file_entry->inode->i_size = MINIFS_INODE(file_entry->inode)->size;
    spin_lock(&sbi->lock);
    list_add(&file_entry->list, &sbi->files);
    spin_unlock(&sbi->lock);

    file_entry = kzalloc(sizeof(struct minifs_file), GFP_KERNEL);
    if (!file_entry)
        return -ENOMEM;
    file_entry->name = kstrdup("version.txt", GFP_KERNEL);
    file_entry->inode = minifs_get_inode(sb, sbi->next_ino++, S_IFREG | 0644);
    if (IS_ERR(file_entry->inode)) {
        kfree(file_entry->name);
        kfree(file_entry);
        return PTR_ERR(file_entry->inode);
    }
    MINIFS_INODE(file_entry->inode)->data = kstrdup("MinFS v1.0\n", GFP_KERNEL);
    MINIFS_INODE(file_entry->inode)->size = strlen("MinFS v1.0\n");
    file_entry->inode->i_size = MINIFS_INODE(file_entry->inode)->size;
    spin_lock(&sbi->lock);
    list_add(&file_entry->list, &sbi->files);
    spin_unlock(&sbi->lock);

    return 0;
}

static struct dentry *minifs_mount(struct file_system_type *fs_type, int flags,
                                  const char *dev_name, void *data)
{
    return mount_bdev(fs_type, flags, dev_name, data, minifs_fill_super);
}

static void minifs_kill_sb(struct super_block *sb)
{
    struct minifs_sb_info *sbi = sb->s_fs_info;
    struct minifs_file *file_entry, *tmp;

    spin_lock(&sbi->lock);
    list_for_each_entry_safe(file_entry, tmp, &sbi->files, list) {
        list_del(&file_entry->list);
        kfree(MINIFS_INODE(file_entry->inode)->data);
        iput(file_entry->inode);
        kfree(file_entry->name);
        kfree(file_entry);
    }
    spin_unlock(&sbi->lock);

    kfree(sb->s_fs_info);
    kill_block_super(sb);
}

/* 文件系统类型 */
static struct file_system_type minifs_type = {
    .owner = THIS_MODULE,
    .name = MINIFS_NAME,
    .mount = minifs_mount,
    .kill_sb = minifs_kill_sb,
    .fs_flags = FS_REQUIRES_DEV,
};

/* Inode 操作：查找文件 */
static struct dentry *minifs_lookup(struct inode *dir, struct dentry *dentry,
                                    unsigned int flags)
{
    struct super_block *sb = dir->i_sb;
    struct minifs_sb_info *sbi = sb->s_fs_info;
    struct minifs_file *file_entry;
    struct inode *inode = NULL;

    spin_lock(&sbi->lock);
    list_for_each_entry(file_entry, &sbi->files, list) {
        if (strcmp(dentry->d_name.name, file_entry->name) == 0) {
            inode = file_entry->inode;
            ihold(inode); /* 增加引用计数 */
            break;
        }
    }
    spin_unlock(&sbi->lock);

    d_add(dentry, inode);
    return NULL;
}

/* 创建文件 */
static int minifs_create(struct inode *dir, struct dentry *dentry, umode_t mode, bool excl)
{
    struct super_block *sb = dir->i_sb;
    struct minifs_sb_info *sbi = sb->s_fs_info;
    struct minifs_file *file_entry;
    struct inode *inode;

    /* 检查文件是否已存在 */
    spin_lock(&sbi->lock);
    list_for_each_entry(file_entry, &sbi->files, list) {
        if (strcmp(dentry->d_name.name, file_entry->name) == 0) {
            spin_unlock(&sbi->lock);
            return -EEXIST;
        }
    }

    /* 创建新 inode */
    inode = minifs_get_inode(sb, sbi->next_ino++, S_IFREG | mode);
    if (IS_ERR(inode)) {
        spin_unlock(&sbi->lock);
        return PTR_ERR(inode);
    }

    /* 创建文件元数据 */
    file_entry = kzalloc(sizeof(struct minifs_file), GFP_KERNEL);
    if (!file_entry) {
        iput(inode);
        spin_unlock(&sbi->lock);
        return -ENOMEM;
    }

    file_entry->name = kstrdup(dentry->d_name.name, GFP_KERNEL);
    if (!file_entry->name) {
        kfree(file_entry);
        iput(inode);
        spin_unlock(&sbi->lock);
        return -ENOMEM;
    }

    file_entry->inode = inode;
    ihold(inode); /* 增加 inode 引用计数 */
    list_add(&file_entry->list, &sbi->files);
    spin_unlock(&sbi->lock);

    d_add(dentry, inode);
    dir->i_mtime = dir->i_ctime = current_time(dir);
    return 0;
}

/* 删除文件 */
static int minifs_unlink(struct inode *dir, struct dentry *dentry)
{
    struct super_block *sb = dir->i_sb;
    struct minifs_sb_info *sbi = sb->s_fs_info;
    struct minifs_file *file_entry, *tmp;
    struct inode *inode = d_inode(dentry);

    if (!inode) {
        printk(KERN_ERR "minifs: unlink: no inode for %s\n", dentry->d_name.name);
        return -ENOENT;
    }

    spin_lock(&sbi->lock);
    list_for_each_entry_safe(file_entry, tmp, &sbi->files, list) {
        if (strcmp(dentry->d_name.name, file_entry->name) == 0) {
            list_del(&file_entry->list);
            spin_unlock(&sbi->lock);

            /* 释放资源 */
            if (MINIFS_INODE(file_entry->inode)->data)
                kfree(MINIFS_INODE(file_entry->inode)->data);
            kfree(file_entry->name);
            kfree(file_entry);

            /* 更新目录时间 */
            dir->i_mtime = dir->i_ctime = current_time(dir);

            /* 释放 inode（由 VFS 管理） */
            d_drop(dentry); /* 从缓存中移除 dentry */
            return 0;
        }
    }
    spin_unlock(&sbi->lock);

    printk(KERN_ERR "minifs: unlink: file %s not found\n", dentry->d_name.name);
    return -ENOENT;
}

/* 模块初始化和清理 */
static int __init minifs_init(void)
{
    return register_filesystem(&minifs_type);
}

static void __exit minifs_exit(void)
{
    unregister_filesystem(&minifs_type);
}

module_init(minifs_init);
module_exit(minifs_exit);

MODULE_AUTHOR("Grok");
MODULE_DESCRIPTION("Minimal File System with CRUD for Linux 5.10");
MODULE_LICENSE("GPL");

---

编译和测试步骤

1. 更新代码：

2. 将 minifs/v2/minifs.c 替换为上述代码。

3. 确认 Makefile：

4. 确保 Makefile 正确：

   obj-m += minifs.o
   
   all:
       make -C /lib/modules/$(shell uname -r)/build M=$(PWD) modules
   
   clean:
       make -C /lib/modules/$(shell uname -r)/build M=$(PWD) clean
   

5. 清理旧构建：

make clean

1. 编译模块：

make

1. 加载模块：

sudo insmod minifs.ko

1. 创建循环设备并挂载：

sudo mkdir /mnt/minifs
sudo dd if=/dev/zero of=/tmp/minifs.img bs=1M count=10
sudo losetup /dev/loop0 /tmp/minifs.img
sudo mount -t minifs /dev/loop0 /mnt/minifs

1. 测试增删改查：

2. 查（Read）：

   ls /mnt/minifs
   cat /mnt/minifs/hello.txt
   cat /mnt/minifs/version.txt
   

   预期输出：

   hello.txt  version.txt
   Hello, World!
   MinFS v1.0
   

3. 增（Create）：

   sudo touch /mnt/minifs/newfile.txt
   ls /mnt/minifs
   

   预期输出：

   hello.txt  newfile.txt  version.txt
   

4. 改（Write）：

   echo "Test content" | sudo tee /mnt/minifs/newfile.txt
   cat /mnt/minifs/newfile.txt
   

   预期输出：

   Test content
   

5. 删（Delete）：

   sudo rm /mnt/minifs/newfile.txt
   ls /mnt/minifs
   

   预期输出：

   hello.txt  version.txt
   

6. 卸载和清理：

sudo umount /mnt/minifs
sudo losetup -d /dev/loop0
sudo rmmod minifs

---

调试建议

如果段错误仍发生：

1. 检查内核日志：

dmesg | tail -n 50

查找与 minifs 相关的错误或崩溃信息。

1. 验证挂载：

mount | grep minifs

1. 检查模块引用计数：

lsmod | grep minifs

如果引用计数非零，确保卸载挂载点和 loop 设备：

sudo umount /mnt/minifs
sudo losetup -d /dev/loop0

1. 启用更多日志： 在 minifs_unlink 中添加详细日志：

printk(KERN_INFO "minifs: unlink: processing %s, inode=%p\n", dentry->d_name.name, inode);

---

附加说明

• 功能完整：文件系统支持增删改查，数据存储在内存中，适合学习和测试。
• 限制：仍为内存文件系统，卸载后数据丢失。如需持久化，可扩展到块设备存储。

---

💬 评论